Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Computational Study of Li2SnO3 and Li2SnS3*
Jason Howard advisor Natalie Holzwarth
Department of Physics, Wake Forest University, Winston-Salem, NC USA
Li2SnO3 Li2SnS3
*Supported by NSF grants DMR-1105485 and 1507942 and WFU’s Deac computer facility
Li Sn O S
Background Li2SnO3 and Li2SnS3
• Close packed layered structures
• AB stacked
• Monoclinic Space group 15 C2/c (#15 in the International Table for Crystallography)
• Oxide material studied as possible anode material 1990’s-present(Courtney & Dahn, JES 144, 2943 (1997)) (Zhang et al., J. Alloy Compd. 415, 229 (2006))(Wang et al., Surf. Interface Anal. 45, 1297 (2013))
• Sulfide studied as possible electrolyte 2014(Brant et al., CM 27, 189 (2014))
• ionic conductivity, ������� 1.6 ∗ 10�� �
��373 �
����� 2 − 3 ∗ 10�� �
��(563 �)
Wang et al., Surf. Interface Anal. 45, 1297 (2013))
Li2SnO3
Overview
• Activation energy for lithium ion migration
vacancy migration and kick-out mechanism
• Interstitials
• Interfacing with lithium
• Lithiation of bulk –change in volume and cell dimensions
energetics
Computational methods• Density functional theory with LDA• PAW formalism using datasets generated with ATOMPAW code (Holzwarth et al. CPC 135, 329 (2001))
http://pwpaw.wfu.edu
• Electronic structure calculations performed using QUANTUM ESPRESSO . (Giannozzi et al. JPCM 21, 394402 (2009); http://www.quantum-espresso.org, Gonze et al., CPC 180, 2582 (2009)); http://www.abinit.org
• Plane wave expansion for wave functions withBrillouin zone integration mesh of 0.003 bohr-3 |� + � |� ≤ 64 Ry
• Ion migration estimated with Nudged Elastic Band (NEB) method. (Hinkleman et al. JCP 113, 9901 & 9978 (2000))
• Visualization software: Xcrysden, VESTA
• Plotting xmgrace, gnuplot
Calculating Ea via vacancy mechanism
• Choose paths of migration:a-path: A – B – C –D b-bath: E – C – F c-path: G – D – H – I
• Relax vacancy structures
• Run NEB calculations between images
a b c df
ed
g
h
i
NEB results, extracting Em
Em = 0.61eVLi2SnS3
(oxide) (sulfide)
a-path b-path
Em = 0.28eVLi2SnO3
c-path
kick-out mechanism for Li-ion conduction
Interstitials vacancy interstitial pair defect
For the vacancy interstitial defect������� – �������� = ����������
1.25�� (Li2SnO3)
0.96�� (Li2SnS3)
NEB results, Em for kick out
oxide
sulfide
0.22eV
0.14eV
Li2SnO3 Em Em + �
�∗Ef
vacancy 0.28 0.91
Kick-out 0.14 0.77
Li2SnS3 Em Em + �
�∗Ef
vacancy 0.61 1.07
Kick-out 0.22 0.68
• ExperimentL.P.Teo, et al. Ionics (2012) 18:655-665
0.69 - 0.91 eV
• Experiment J. A. Brant, et al.. ACSJCa|JCA10.0.1465/W
0.59 eV
�� ≤ �� ≤ �� +�
�∗ �� �� = 1.25 (oxide) , 0.96(sulfide)
Overview
• Activation energy for lithium ion migration
vacancy migration and kick-out mechanism
• Interstitials
• Interfacing with lithium
• Lithiation of bulk –change in volume and cell dimensions
energetics
• Modeling interfaces with Li help to understand behavior with excess Li at surface
• Can provide insight in to electrolyte or anode functionality
• As electrolyte Li2SnS3 would need to have stable interface with Li
• As proposed anode, Li interface calculations are to help understanddecomposition process
Interfacing with lithium
• Preparation - surface vacum calculation
Six extra lithium left on each surface, 24*(Li2SnO3) + 12*Li
Li2SnO3 Vacuum Volume ≈4500 bohr^3
≈ 19 bohr
Li2SnS3 Vacuum Volume ≈8300 bohr^3
≈ 23 bohr
Li2SnO3 - Li interfaces
VC-relax, cell-do free “Z”
-14716.9205 Ry, ref 0 eV
-14716.8014 Ry, + 1.62 eV
MD 300k 38 fs → VC-relax, cell-do free “Z”
Li2SnS3 -Li interfaces
Unconverged
Overview
• Activation energy for lithium ion migration
vacancy migration and kick-out mechanism
• Interstitials
• Interfacing with lithium
• Lithiation of bulk –change in volume and cell dimensions
energetics
• Goal is to better understand decomposition process for Li2SnO3
Lithiation of Li2SnO3
(Zhang et al., J. Alloy Compd. 415, 229 (2006))
• Li2SnO3 has been studied as electrode (Courtney, Zhang, Wang)
1. Li2SnO3 + 4*Li 3*Li2O + Sn 2. 3*Li2O + Sn + 4.4*Li 3*Li2O + Li4.4Sn
• First cycle irreversible
• Decomposition Occurs between 0.4 – 1.75 Li inserted per Li2SnO3
(Zhang et al., J. Alloy Compd. 415, 229 (2006))
• Residual lithium oxide matrix said to stabilize further cycles
• I studied beginning of decomposition
Modeling Lithiation up to 1 Li inserted per formula unit Li2SnO3
• Assumption – Li will populate interstitial points first
• Run VC-relax no symmetry = true for 1-16 interstitials filled
• Choose interstitials to fill via random number generator(random.org)
• Run multiple trials avg data
• Interpret data from energetics, %volume change, change in cell dimensions
Unrelaxed structuresAll 16 interstitials filled
Volume expansion change in cell-dimensionsCell-dimensions
a axis
c axis
b axis
15 Liinserted
Energetics
Energetics
• For each calculation I can write the process as
16 ∗ ������� + � ∗ �� → ���[16 ∗ �������], � being the # Li inserted per supercell
I analyzed ∆�
�=
�(���[��∗�������]) ��(��∗������� � �∗��)
�
This shows the insertion process is favorable
And is becoming more so as the #Li inserted
Increases
∆� ��
(eV)
Conculsions
• Kick out mechanism is likely mechanism for Li-ion conduction
• Experimental samples (Brant, Teo) measured for Ea not likely to have significant populations of native vacancy interstitial defects
• Li2SnO3 – Li interfaces appear to be stable
• Li2SnS3 – Li interfaces appear to be unstable
• Bulk Li2SnO3 becomes semi-unstable at ≈0.5 Li per formula unit
• Lithiation of Li2SnO3 becomes more favorable as the #Li inserted increases
Acknolwedgements
The computational portion of this work was supported by NSF grant DMR-1105485 and 1507942. Computations were performed on the
Wake Forest University DEAC cluster, a centrally managed resource with support provided in part by the University.
References
• I.A.Courtney, et al. “Electrochemical and Situ X-Ray Diffraction studies of Reaction of Lithiumwith Tin Oxide composites”. J. Electrochem. Soc, Vol. 144, No.6, June 1997
• J. A. Brant, et al. “Fast Lihtium Ion Conduction in Li2SnS3 : Synthesis, Physicochemical Characterization, andElectronic Structure”. ACSJCa| JCA10.0.1465/W
• L.P.Teo, et al. “Conductivity and dielectric studies of Li2SnO3”. Ionics (2012) 18:655-665
• D.W.Zhang, et al. “Li2SnO3 derived secondary Li-Sn alloy electrode for lithium-ion batteries” . Jorunal of Alloys and Compounds 415 (2006)229-233
• Q.Wang , et al. “Preparation of Li2SnO3 and its application in lithium-ion batteries”. Surf. Interface Anal, 2013, 45, 1297-1303
• M.K. Aydinol, K. Cho, et al. “Ab initio study of lithium intercalation in metal oxides and metal dichalcogenides”. Phys, Rev B. vol 56, no. 3 july 15 1997
Density of States Li2SnO3 interfaces
VC-relax
md -> vc-relax
Ef 3.3eV
Ef 3.1eV
Li X5
Sn
O
bulk
Vacum surface
Ef 3.4eV
Relating energetics to experiment• In analyzing my data I found a relationship to a first order approximation
to the experimental Voltage vs capacity slope
(Zhang et al., J. Alloy Compd. 415, 229 (2006))
∆� ��(e
V)
If I subtract the y-intercepts to the fits above I get a fit representing an approximation to the change in Voltage as function of Li inserted
My curve appears to approximate the change in voltage as function of Li inserted for this experiment
Vacum extra six on surfacesBulk
Li X 5
Sn
S
Ef 1eV
θ(b-c)
θ(a-b)
θ(a-c)